# 开始运行 LiteRT 模型您可以通过设置先决条件并下载所需文件，在 Qualcomm Linux 开发套件上运行 LiteRT 模型。使用预编译的 gst-ai-classification 或原生 label\_image 示例应用程序运行 LiteRT 分类模型。 ## 运行 LiteRT 模型的先决条件在开始运行 LiteRT 模型之前，请执行以下操作执行以下操作来设置基础架构： 1. 设置 Qualcomm Linux 开发套件。相关说明，请参见以下内容： - QCS6490/QCS5430：[Qualcomm Dragonwing RB3 Gen 2 快速入门指南](https://docs.qualcomm.com/bundle/publicresource/topics/80-70020-253) - IQ-9075: [Qualcomm Dragonwing IQ-9075 Evaluation Kit Quick Start Guide](https://docs.qualcomm.com/bundle/publicresource/topics/80-70020-263/) - QCS8275：[Qualcomm IQ-8 Beta Evaluation Kit Quick Start Guide](https://docs.qualcomm.com/bundle/80-70017-263/resource/80-70017-263_REV_AA_Qualcomm_IQ-8_Beta_Evaluation_KitQuick_Start_Guide.pdf) - IQ-615: [Qualcomm Dragonwing IQ-615 Beta Evaluation Kit Quick Start Guide](https://docs.qualcomm.com/bundle/80-70020-293/resource/80-70020-293_REV_AA_Qualcomm_Dragonwing_IQ-615_Beta_Evaluation_Kit_Quick_Start_Guide.pdf) Note IQ-615 仅支持使用 CPU delegate 进行推理。 Note Qualcomm Dragonwing™ IQ-9075、QCS8275 和 Qualcomm Dragonwing™ IQ-615 快速入门指南仅供授权用户使用。要升级您的访问权限，请前往 [www.qualcomm.com/support/working-with-qualcomm](https://www.qualcomm.com/support/working-with-qualcomm)。 2. 使用 HDMI 将 Qualcomm Linux 开发套件连接到显示器。 3. 将 Qualcomm Linux 开发套件升级到最新的可用软件版本。有关说明，请参见[下载平台 eSDK](https://docs.qualcomm.com/bundle/publicresource/topics/80-70020-254/how_to.html#download-the-platform-esdk)。 4. 将镜像刷写到设备。有关说明，请参阅[刷写软件镜像](https://docs.qualcomm.com/bundle/publicresource/topics/80-70020-254/flash_images.html)。 ### 后续步骤 - [使用基于 GStreamer 的示例应用程序运行 LiteRT 模型](https://docs.qualcomm.com/doc/80-70020-54SC/topic/getting-started.html#run-a-tensorflow-lite-model-using-the-gstreamer-based-qim-sdk) - [使用原生 LiteRT 示例应用程序运行 LiteRT 模型](https://docs.qualcomm.com/doc/80-70020-54SC/topic/getting-started.html#run-a-tensorflow-lite-model-using-a-native-tensorflow-lite-sample-application) ## 使用基于 GStreamer 的 Qualcomm IM SDK 运行 LiteRT 模型 Qualcomm Linux 开发套件包含预编译的 LiteRT 示例应用，用于运行 LiteRT 示例模型。 gst-ai-classification 示例应用使用 Qualcomm IM SDK 插件，在 Qualcomm Linux 开发套件上运行 LiteRT 分类模型。该示例应用程序使用 LiteRT delegate 实现硬件加速。 Note IQ-615 不支持使用 Qualcomm IM SDK 运行 LiteRT 模型。下图展示了 pipeline，它从摄像头接收视频流，进行预处理，在 AI 硬件上运行推理，然后显示结果： **图：使用 Qualcomm IM SDK 运行 LiteRT 模型的工作流程** Gst-ai-classification 示例程序执行以下操作： 1. 打开 Qualcomm Linux 开发套件上的 IMX577 摄像头，并设置特定的分辨率和帧率，例如 1080p 分辨率，30 fps 2. 对每个摄像头帧进行预处理，以向分类模型提供输入数据。例如，gst-ai-classification 示例程序： 1. 将 1080p 帧缩小到 224x224 分辨率 2. 根据模型要求对输入帧进行归一化处理 3. qtimltflite Qualcomm IM SDK 插件是基于 LiteRT C++ API 构建的，可执行以下操作： 1. 加载示例 LiteRT 分类模型 2. 使用硬件加速对模型进行推理 4. 对推理的输出进行后处理，即在输出张量中提取预测概率最高的标签。 5. 将推理结果叠加在原始摄像头输入图像上，并将其显示在连接的显示器上。如需使用 gst-ai-classification 示例应用程序运行预编译的 LiteRT 模型，设备上必须有 LiteRT 模型及其标签文件。如需下载示例 LiteRT 模型和相应的标签文件，然后将它们复制到设备，请执行以下操作： 1. 转到 [Qualcomm^®^ AI Hub](https://aihub.qualcomm.com/models/inception_v3?searchTerm=ince&chipsets=qualcomm-qcs6490-proxy)，下载 Inception-v3 模型。 ![../../_images/download--copy-sample-model.png](data:image/png;base64,UklGRg5cAABXRUJQVlA4IAJcAABwVwKdASq/BB8DPwF+t1arJ6uuIlSJWcAgCWdu/EsZT502irFEraz2lHo3Ii4+nREF30+/9Xp/eo3zAecl6av8T0sP/09l7+q/+72M/PK9aX/k9IB///b/6W/rL/j/Tb8d/ZP89/hvx89P/yT6z/V+nN9p5a/Tf8DzQ/lX4E/mf4r0P/7H+H8e/kR/uf4P2CPzn+oeC3/dfad5YWx/8X/wf5z2Bfb/7z/2/M1+C/6n+a9Uv07/Ifs5/kvkA/ovn5/yvCg9E/bf4CP6X/sf2r91z/D//XmS/TP+D+VnyH/sj/++zQKHi+9Y3esaJSGhRMAnRgTowJ0YE6MCdFzCY0Iu4ksWXfh9UcXcSWLLvw+qOLuIzOpSjtpxJ5DerJ7BXkCwlwbkVXM2L97L2W5H1aqIfOzpT0pm/2QNzADj7f7MjnVUKLPtGIJXtbewXRX2v+uzDVpxN9gF2gg2ct9hW+wrfYVvsK3LxdlxcNUzLPW57GHSsz2MOlZnsYdKzPYw0WvzstU9PVXkyy78PqjiUIJzo4TFHwNRpg2CSBTtfsLZbsA76Hveqee+1gMMtMOBRNjcuaeVoLoi/GZOe6HkO2if5kVnT0/ayn5KIuXC/OKkSWLVmTiHEPviCQmIJGKvV8q9BhurVtuluUBZE+HZr2GBOjAnRgTowJilwX15mwyUGGmdiR1RlItn/aRx5F/N699L0w1Esm7I2D2+RS/qbUjQH+LdAEhlh7gwz5tYIRe5lTrpumSqJcXtVKpXOu9OnFJATowJ0YE5jH24C/YMHtXbhmcWD3u3VIalR9ZV+FAGPKGVO12AICdGCqI44IARy+u5CoRHfOKu9XKY0NAsmrvl0pIPQ47wcd4ONV3+0MV43afCA4tESVC3iI3lq/rrl0OjID0/6BBGgTtyTmWVsRpz4xjC8sQdn8C9BJZi0RPdIGoDoIVxDwfArBIsuKNH+7PqQK071EJJ8FYLGgRFgDfwAzZS+EK+rnhT8L51WegLSBiOEcdbuSYGEhIHZaCkSC92Z/Zw4vA0yIdYwCjjlcNO4AZ6uaRfd2a5/NPC3t8dCnko5ju/954BasOQHpzowp8JjKKrVLbuuTbFadFLQllNRCBtoztKxrmyWe+IoFjc8zuBmMBc8t2ggvqINTFZ+XmHcN5qfz/11E7vpocDxRZ0ZpLVqXukUkBOiRSbbPCKQQQxFHxDvNmm4ZkINEvVTWnXZYHMsoSqsj3ZjpEGII0deCD+K5FS+ogCpcYht6eSXZbN7aPlb4qBwqBpGs8fR+TpxSQE6JFJtCsjsu2TwgoJPmYxJ1VWeMje19ptRPDvLQnJGfZIwSWN102LLvw+odQZMU9gfCilFrR1HDzsm63CzvNbndI7XB8yLQZWSuCZJcRBncH5rLgxmYq5CCoFjktl463mRJq5S+6z/E55R8/xfO4xdnCKsUCSmWfnF5k6tl2DDpMb2KRyhDo7dKHmn/6bJ1IC0X1r8EA4Hw/g47wcarv+UuH66LGlWysY1AJJz2Jx2UnxClVo12vflTIkqc6cLhjgP85CZYJ3p6rN43Xo6rW/7SMcFp0mMkJWbpc9SzXFvnmK0LwFFSI2bz1A7UwB33hx6sS7o0JYhUhSr8hw2P9IV8z38R1LE/KKYoIQYNGq752KlydVCZOcdB/qt7wGPq+P6n2Fb7Cr+P1k6s6U5fc5kUE1LrkJ4HEKmuyoC3umXhDZh3TlyveUAyWnFQgA6k6Yd60DQH1vtiWoIMdFxTEz4wuuICoNZUJDtbKjFLYROA2SzXSXl4zxk6MFURwBClXgap5ykuvLxEE2+g/w8w2zPXvcR8n3h3oSNF+6yWbAjE6aecjB63lFw2v0X2LDbroBfOZEF2cqCsqUVkNp3ZX92Q29FwF9P4X/jq28Kjb3ttAmwcUHySbmej0kRPBzk+QV8j15kzN/v0uixNPWKzoQrH0XPe/Y0dHRycmfC2tQSqSp1dNm71jd4wCY+rUR8BuLqRKzIB5yN8BdvEyKwwEPYb7NIj93li6k43yAR90k89gsqnr8nam+sZqifXY+GeOBXaacmkfW7Ty+DnEsRO9QkXLdPfgvVeokGJsF222E69OHXeoBCKIjviumYSY48ykGNv5Gj2WhrWRgKWp2eNLOxrFFCoLjoOG+DZW2+QDEcwzv6iEdOgvseetdlViqJ7Qyo2i5sESB1yFKkf8tDr6i+i62uArepdy0pcdpEyA1ClDg41XcyGtUzd8ENLCCRjadfAMoYFEFvt7r9XJwJQU1jNE2nsudX/4nsX6IzbqW5cconne/5vCSRKE77R3zWyGyVtX1HxBCgf6oweUzkKuRkZpEzQrrxG7AGWQqO/+MFy9CPmjv/iWFlbR4AImshRc66r9iCSGZYBrOMOO8G8OCXiG7mVzTpCOr4SD3Vz1SRLt3EyHH+YCWfhTV32V+ihUQeEWQE55K/9q8hZ4WU9y+nnuLqEkiT9tGAd5/ldq79plMUyLt/D6o4uljpCiu99mYBOKSAnRgUleQ6LG71jd6xr6S6sBUlggrys0PB5vhWemYqfx2ls1/qM+YrUr6xmXnY2gspCfuCX3e7SSvToY8bD0XkUHGY4sNgbjEUoJ+tUx1iN6zxGsXiID1vg9xJYr7diSXauJEwTsQclMII0EaOvNbkA/pggDOaFFjd6y2NGBOjAnRgL2HWc45y+mikmlHMTRiLFhUlTZ74yMqrTz8KcJy+Zhy0H9kt3PlLIMbJH+UIpTfkN3rZ+GNeEEov0eSwmxftpsOkxr+LPgqoVpybG71jwDx5QNOIcIJRAQmAgEkbwo2+94APdYHVy0VnmmOCPdgORMWkelF3mOkn7gsLg3DcILexZ3mWhVoJ0YE6MCdGBOjAnRIpNsGHD2MTvJ1EXKNw8qAJpdTL9dCeI5Pz5VH7fo9RVYAiLW/soPlfK0tp4UwCpYhVnEsA6HJIxjRABM2fg2MYgcxQPdO7fjqF5lZUm3Juy1Mmfa9c0hMsU5fyNqGpmB9zAAL238ZCSMRo9bz+bvWN3rG71jd6xu8YBMro1pDjtQvg6B+Q6JZF3IKm1TWR+LB6W5fspiwFxgZxu7zU6OqKOVItx5ZtF6IR3WQ9x2CdV5uwOYvOgYlc7dOG6AXo2GkGy4gwzEbXgQ52kxDkkexBMMCdGBOjAnRgTowF7DrIn/qfKe5zr/k2gVJe5laquM3rQMvvuwiUgeI/HW+3DK2U6M7VfyyIraRd+H1RxdvWNPCw2CQW/qhhNUGJIuUkBOjAnRgTowJ0Xp45C4Adld25lzZ3aHejlKg/Sospl/jCxP6MRaCuc8yB7Gz5bgnsOixu9VYRkr56fesbvWN3rG71jd6xu9Y3esa+kuqqcQ64dIZksOfbbikgsgc9BpGaOMSo9qHd7iC1CvABh6kk+jOAjRNHeD0WL0BOjAnRgTowJ0YE6MCdGBOjAXsOvTX+VKRdugbiLYWPdq8t2nmuUpufWcrLAI1dNDsEz+B895ST6I99hW+wrfYVvsK32Fb7Ct9hW+vLTIAbpgvVfG8JHWm842xg8dea3PYrqVmexY62wxm/ukOQgZH45G9iTv2Fwxc732Fb7Ct9hW+wrfYVvsK32Fb67VVHPgF2/mT+CTUJf1QwmqDEk3xTdkbB7h00gAH3cdrbYjd6xu9Y3esbvWN3rG71jd6xu1CdUqQc1/Dfw+qOLuJLFl34fVHF3ElG9OO7wcd4OO8HHeDjvBx3g47wcd4OO8/0WN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xudEJXfAhbRsu/D6o4u4ksWXfh9UcXcSWLLvw+qOLuJLFl34fVHF3Eliy78Pqji7iSxZbst6mC1m71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN3rG71jd6xvAadKfcqRM3UAmBpICdGBOjAnRgTowJ0YE6MCdGBOjAnRgTowJ0YE6UwHEqACh+Dd0ykG/B6O3fhkT1yUtI6OAc3yaf0ghfAquHr5ziARvkO24hwUfDRMb0NfxhERykA+RvT71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbvWN4DUsAeEmGPEP+B3/yyRkOR/nl/h9UcXcSWLLvw+qOLuJLFl34fVHF3Eliy78Pqji7iSxZd9tDs+RzItvWppkxzaT7H9U58ZZCTu6X1RwTxJYsu/D6o4u4ksWXfh87xTnbZNxJURHUcN/D6o4u3v+SjmvN2RIBCvs46+GvccIA47Kz1HHNPHd0gEYIYdFjd6xu9Y3esbvWdGo0EGmSWi1lUCksWoJ4kyy3OnsEkigGQKyKFcbjzVm2AG2f+faPrA7UzYwj/ABwyZw20ZYs7Xh48h1p1MpmqfnmX20mqpbY5i8jzFI1T2il5YoP8xnQRTBiwIQwO0Xs5LccAGe+uHiVOc4OmB48TRunxiBaFeEvKHLVtLMfZ3ERc5e1q19LVPma9WXZi68jqQeTTBzkC71jvi9W+xATsd7nnkoKZ0ZvqIUG6IKNLJR0cDSREcBkgBwELG8BXnGpIpJ4odb+oKmlwYXvLgGekc6WNWlTBRRKZqFx+oQHA2wCHwQwTkG1H+AzaNuQ9xgS/QBSOS2GbmdwnydZ3ArIbYAqmEIWdBMpLxcY28GPYwYKx7i7kpJmSb3EvNTZpr8SW2oAZtsXrEud2t2grhLSM1V+LZGQCVjLMBdpQuwFHKeNPbBAREVt0jJ8oN700nXL7LBzG3aOps7eXB36Puktb8UMJM2VzcSXOvuIKNAzC5DVdKU1yoRZ00gExfMrnBPCKYsQgX4LkE/DkTbNt8d8HODP+7Z5oYPwkxSch6/CvpsY+R7t9pNoeyxAJklk4PfzheOoxQPVzV+z6AXxfSI2P2RCba7To6o0Ru9l+X0ZyDSoEa6lSLDs3hRymMyMIWEsbdjDkYQ/YryW2nmQS/1fMkrTluBy01/XNnUGbE6xH+uoo8eq1cFVfQIJWC7qhw89RT5S3PIXGyTWal/vYG9npRTa0pqTURXym/VgusfnWj999sXN4F2ZmqBqhTRLvAoe5SDjeyKUG8p6cCgMhVwESWlsZJvgngTKGEV6cUkBSQenTkJmhw1kLAdDd+edMN+8gTjDM+It5u+mKiP0+xCzD5J4Z77TvVH1qFr+O2lpEwBZvohWzgIAZO54y6on4KlfaebjomOqB5vTw+H1RxdxJYsu/D6o4u4ieD0h5zN3rO5jobb/e2RX+ojXsnr5MX9ObRlFyW7hdv4fVHF3Eliy78Pqji7iSxXcIer3B0WcGSHV7wvyYSnbBp2azcE3SywwJ0pGM4dGX6qs4KokNgSx2CFA4oj0m4il6C0PIvwmlY6p6zO440x3FKehzjH2x+TwM8L7BbTRX2Fb7Ct9hGAJycI73rgwRNyecsVW5H2MbdlqQ11i9Iw1MO1lTBsHTmfwCHZKuKf4eSbuGniK/fB31I5bR3hv1D8cVakJZpjGvLts8qo8ore8bC6QaekqMYATZmr+7XDjZBMdkbi/dQbhyvN2RsHuJLFlfbuGtnojtwFlTeUODdS9iymlAq5Psk8Ah9HvUqkTbtFP9lhCILG1aQAA//m4zCnj9WYfEUPUnHDuXjqznXw/SDIB04G2q+J5Pj76em+prF7w8DLHcrOcpqDnnnfwa5M7uOZVeSVvLNJ8VvX2okreYZSH1ydluKGFGTDgZhvWd2b47bo8eO40ltcOnJCue7KqAxOgwB3YKKMTyzuXTgJ8C7Q2lwPIBBAYfSbiSxZd+H1RR6vcvvLS9naF3zFiYE9VpuiFhuyxxPl04CXesbvWOgwwJ0YE6MCdGBOjAnRgTrYShz/RY1nx2Yt80Ro681uevsvPCbony6cBLvWN3rG71jd6xu9Y3esbvWN3rHQYYE6MCdEGbMmCrO+d+T8NxpTpO+mKhQrJxq4z2jshz6Aw41JFIkBOjAnRgTowJ0YE6MCdGBOjAnRgTowJ0YCxM/eoBSmuwVfhJuEKuuC0M9gniqPiCANobXAp7zarlYdBHxVEg5HMm8hW/KSvHahBL0IS1LzLYToK5erfYVvsK32Fb7Ct9hW+wrfYVvsK32Fb7Ct8ryrIXePTxzELMxpyw5Byo3NWQDHbo9+vTe8NSQE6MCdGBOjAnRgTowJ0YE6MCdGBOjAnRiheZh1ZnsYdKzPYwzpyMf/LPRetOXrG71jd6xu9Y3esbvWN3rG71jd6xu9Y3esbxDB7iSxYsJy04zyXPwJRfh9UcXcSWLLvw+qOLuJLFl34fVHF3Eliy78Pqji7iSxZd+H1RxdxJYsxgNF5RfgAAAP7+Ld4OzfMjZVjlAAgt2po0NAAAPl4rvclJbloKWeiiBytjC7eaO0q4zLP4kEmHJSZB6EGbwpjziF6YcG109R+v3jDWp+G+d/VK/4B5DhjWg8f2FXlaY8+hc74tu2s4tz8p1boKAfle5C2B2DOv7KcTHmn5l6qWJgQ/nVVi9nMTkfkcJadZ2sZy0q/LoEfiTArh/wh/Q/N4cIBk0sTzgOOOOG2M3dS7Lpgf2Z6z04Gv9xJLyv7kgso6QnQUCHr6wU8lIwjWsXvKbqAWx9h9vTKvcD8/+M5CBzV59XlCb+emNUgzygVJNZTEXewqZWGD03UjwLbdemVwzyE2VZgFzYn3trYMfw3zSHyL/BqYp0Hrznh2e9wiaeroEdRxUmt5Xxp8wQZqDV8StdVnwNCCcr9k+jC87ndwoXEU4m3NiDte18ct9uHN8w6IYvsyDabaRQmSlI2c1GuPp+bDi8Hc9vcWadSCLA4SmjfzUN6oXKc7j9t4QMMSQhVQPaEaYd+uXEKDx062x63hS9GDP5Co4F65LjBa2ufWrcYpgMAiWI93WuiRKduglYYaPyVjxlC2zOpduL4TyYiPk+8V+bqMQghAuldpnTvzw3+syZaJRh528km4JKcU1E/70iRlciATuM+45b5QbCyRqGcM1Deig5EHawH2D03hpZXZB55omABnPmonvP/wC8x6n1ezHS0fBF9VqAD71LN7KfA8LM6IlNOGBshH7vW3yt0SEYVVhVWFVXoUaEVXBVcFVwVXBVcFVwVXBVcFSzul/k/baQWeR/7V4fVXBmIfE4kWic2u8UK49EY22wd/up205sPoSHurFqwMr5E78OUgyBHll7eEryv5zgxxqxKY7nzjsav4FXk8kcR2WZIR/YxwHElM5DD7vVKTR5Aj9KaU8xWlgaIsJG9E7kt32wI6DcOVXkGjSnsS47yACIajhxWwfzk8ExVbr/XIJ9PLe1hKtwFPH/ZwmHGRzP0mGtFqgI2QMr+LAQo/xOhK2dPWg8C578Apkvaiy2cRYKdWlFIDpS/+6S7qTEndKhQjObOJ8c6V0DTnYIlyMkpAMhveB+vdaKxbMhAyrqS06mpZj3VJ8kzLfly6f7qMVDDCpjaDzOTtRDztYX/IJP4Hfj+Bsy9gM+wNUZOk8zPiKkxmC11xog3FJ59Gg30FM86265qyk8lLUSR8jkQCmdTISf8gWwSKsLKiijblNFDZlwXtN2f5dUfDY7Abxk9mM2VVY2/vvf+Az3ySt+OUF4DI0BjdMsIB4HuwQWwEXz5nVgfSMpj8yFu6NyomvA7CuQOGvuWIbAA5rvo5cgaYuSOlufuEqy74U9yneL0/KFQyTIjXXkDb5/YbbSP4ho9g8S3dvqM1YagIIl39p3BLkWUFjlCD9UZ4pEXjpRUWHLSkETDQcALg6PtYkNWAfZ70RGo9xjRsp0WEdrWHNNJNsQFTBrHhYu0B8E+AiI5A6QxVXfrwh7K/bdf5DYolgFzHh1yBnEjEhYsnap6MUtZBJ2Elt/phj3zt3Sb0fsg53/ZDSvrnWBMupNWjiA7ASJK2rJyBbZIcAG37BPBcg8DPk6itfBL0krjCGcHo4zP9QiS4FSvd+fiwCv5rDKQ6i7N/ePmD3v80W1OrDUWsYAfQDtPPrQf+iUNSoHnLhbJ8jODRNhPRd1fLTFNKXU3gFhUINgu4fJ8nLKle30Xkfv4MAUR8YzENAMIBtK44AC7uwFxqXgJfW48AeNiYoL0HOQ3F0n5aLV8EXD5ueO2HOCKZ9Uz9RmxzyUuJBc1olIxRHDJz4C965CeoP8/zqtpgXmo84iys+KozM0VVU6FhYuAv+kUWUMj5o8gGkapNxf7TyrIuDRRc1DADWGp7kDBE/29zhSYAreQbqZvIV4XfMjFuHvt8iRZmKYNL4WZ/bJ0DJQodobC4GpwAbeyvfW5Wsu+zMfvUdw0xwKGSghECLGShDvhp2EwpfiF8DtOTd5la2WeuvdnJLSszvjN6Nyv8jI+4NuMQzrOyRTnPqrtUFplacndfjAUcJq9unrDXFBsCVTu/gFim6YEr22LwzY3gzdvofhgYYsCQ7Q0+aVpwZTuIKtUttrwv6xdXoIpHKJWwtZzIJ0ob3mqx92N1x3O4qyZuLbzYrzlrcOc/fczJrkApHUHaUq2U3/KcR39qCpX1tSmRrpDXagTsO5KkzYHEnO6VbNO1lyH1D/W12ktv1MuBEKiUYGlWlMF/Qf8DXO5jUgUA8vhB6VLxoDrTGLpReMS1oI2ncUpaKngCv5bFVUJ/pxz5jjW5WEpD1orM/BVwSZxziMklkmIXvIVPsd53MVO9ZeEFv7YvPfIbir9wIK34HA3TZjf4qBvTZach2rPk/uw5mWLqhygrr9aeT6ytbq2knJlSdkkY4Jvp1c3pCjdRmw15+VgMSzwjcIKZCA7XXp49hkC7Yy4lt9wmx60xE6yKAraWOoyQJqA8XPFHTrzpjaOWzbMGGLUabeOVK1yhYw8hn+oK4E6R3opvGZ0Wi4fQugxc+AhL0N4t5AyJKOq+KnPZDMWOMDPeW/CLRbvGc4VYby76iJCDgL+8n+R5+42lWpFSKiNEqZA3cu1Z08bhPblMaUdQP+EBOiqHWfleVselfgfHwEpr7e7ck2Ds4r8bR1l3TIxGjM7nAsAcpgUBt26yy9NU0Dq6PzbYF30EJUpXuz8epybgyRDCW/O5OEUmyJXccfDo+mrinmUIvdksXrTK17h/YEJ/7Cu0OxcwJq1H03URPp8iJThzwY0uFJXVhZluXoqeNNU0Hms1fKPxG8PuQyrCJlVJH3xysag1Zjo5oosrtC89z2CQzmaqVfYgkIx+KDHKhzm3FrgIf3Jy/ro9ze7u0oZP8ySOLVrjUv503UQESHriTQUM/His5ThVn26m3o0ior6Lc9X7vEgzeT1Ut0EezU+mL0qwIUb111lwMcM7CQC+eM5AykQWKBpopeejvt2WM2Zp14d/PcC8iN1xsUhS09JPdlwr24qWkydLqp1kuTy2dv4LtGC/sOhfi7U1FC8CwLU2uKqgx5xVc6CasSUFqB0OMuvCVOeBTEqA1OT/H8LuVp6+rYcC9fpvT6K/WNHNZH4dRhnfrIeXaFsLWjxjXWYVDlxj5zdlXoeRXcpodOILIA4IWVaMDj3TWpGLkjk7Ki8fC0TTDtOxZG8/Gl35DnvTgl9SWCotqTUJO3iHK4B0OZPL2RZ/4I4GbRVPe8sinavTQS8968QmbvZJ7xl40HvWqze0BYlMsBWj1M7PFrnVJDuVVEMDfHrOfhiAKBb0zgjsC6AKmQAaOwmr9XPXVfxJLXBk865JOCDeOhvRWtew98CMGgBeT6zD7NnvLT3coAOA9I0LdhSRNlkXykneJhgfo4kdNaV49Z5SPAmvDEm5+kfLx3CUE1k+4jQ6cI1IM3PElaLCVCh1DVYjS1Tks7vauKqK7u6Uyth1yYq6Cr6TGjhS1q+DoBdd/r7ZMdA2p6B/rFkNqNe1GA53qH0G+8tiG8p2gYCjHriNeM6B4fCS9jD00JXGW/082agaudDebmMoL6pAbar1/K6f+47UrqRT7GjuKTWdEmuy9D/m6jOVOdWpdZd3A9PdfwNNYeeBIN0xAehWn+YvsVghyk/OutgmqNo4+zHGO84b8l6fmmCfiCLSsOYHZkY8JwjTHSV8TNz1sTek8Hhd+eezSkLwgvYCDIEU1zO/iyep4GbrQvfR8pguGcd0VkR7nPy8c91L3JneRbCkDVSgNdmp0C0D467LNCv8eMk13kZlTTQ9kUaoqyZbCrm7sHbk3BVH1apPUppD9Bumn+A0dV0AsXukueCm/1NlF5twRYoDgECU90rIQCL6/8JzSqj4Xj6i+cX6OUP1RwNiBL28auL3IBHmODtKZSLZtc6svARW9Z1ZhxoDRwCLsxGSh4yiTuMn+pGL12BkH39qtRQ9OFNNguD4fjXDd4nMGtxnpQDvavzbYxDY9NKnFhOo18y5cuLepqjCKNV9+8NJoLd7hFRa1ucFlCGnKYNSI4fAvehosdnDGxwHCuKzPl7Qm707mjTNdlRIZ78YwnnZNOXnGjSIOqCDwy+zNY0+nmQN2h/OlZAR+vnz8he8wd1WAiJhtYKo/h3MsOTLD8jWwoZDGKl5C4BVW8AKzjkddneUV8DwNnIfARp0OAGItjeVBQ3rXsoqz4H78nm4ZrT276CQZAFBZOujRPGfDPIxaHTJrOogpzZf0qSe74iwLVI5BQaCBpIVaSfazLNPra0cPdnDoLWlZWJsklHX+VsOmdbibB1IQ9v8z/D2J4gcX6+9IoyRaOfHkwCdrFvqLi63KLsDWBby7I3SibcTVLU/aI6T1/obBFqHTt2mfsZHlc9WH7jqJONwsqdaHJmfsZWVI1DfDvE1DnTec4fteGgP7CPn1kbhKuZ9EhyvuNvCv/9V+bR2qinTneHR2bEndsH7JLns1Pl1TEqz6aEmF+RwYWK6tgq2CbH9Wsn8MC3ysLMBS/tynSOPK1lfS6QsdgxfOeRdL/dIvqIPYe1Pwm8DOEQua6tFd79HKaJ0YB8x9XPOrkLsqL+A1LhzOWY/2U9gYS13at7K4sd3PPS6U7wWSbNOTVidOmkZ9EYWvRG2PVGabP8nEKjoc816hsy9Oxr5isVhuzcP/mLSKksNdMY/oZdwb3fEV/tkeKdo0dHwplSw6/Dq4d6O8koFNUGnxlJUPLZ3/8M//T5C9e8H/plpVhYXZP7Vy5RF3fvnkNtcPz7RaNxTXXqkokCnMGXU0BrSJxedO7bFm5RRPsDvRy15hj1aq7gCvMzMOb8q9VNrioUIEaZe902O8AAAMvSrKniUSAAUSLKwAw1N4hZe90/JTX8kiDUpTgKig7PPyjjhVMsEOpybxvu6AEBklpz/G7jCRV5eBD7fZAdlwFz/8iG3sIqI1Awp6N3O39Afctd0BYmi+b6N6l/1Bdyvq0Jn041SMLXndQVvahQ/iKb0fC87ouS82NpL9vgGprEFuT7I8FDXlmFqKvDdaB97FwYPYs7DkYddUu6Zb+n4DAlB+Iy6DCR7Ygi8aOnIPeUCi7rU2fjpZselxeIEzlyK0J8N6+Y9iok4JiEwF6GWYIKz20v1ABlqkh0zx4edZNl6sBut7o35hrPaTcYOxtASBEo/OnsmioTyk4B7zHm6YsXzizg5XZFPVQ2tXbCl9l+9z46aDhYyr03mUNjlmf0PqoFxKaO3Jim5U1CmSzophcTIeXvq1U//K16ZqxFsLB0NzRhdYLggDPmSTZVWhnNyea4PTf3jATSUvNVVTFnGDVyWWe1U1cxQnwPmjy8QCEUEfj348bHF8QnLcdX0thHwCFQ9tR81cr+KpbxR/zKb3s6vfpQOcnkcEostNMZPuxi5lFffsfBmOCFDjxrVx55KS/I0UbAJfIc3RJWW4IVJR9hozJ9+pXbhk4BsOHSaSD1+11zkxmhv4sw4r88PMynxO4bdmVXr6OD6Ow6xot1ykIP0XU5xfbp3W8ufj32yCO0zCymRCaajm6mY0VID/6YWtAdtXkkcxj64df+UMxAjkylPH3jCw941uhETH4QnqO9U4xJZob4/QybKVnp1K3EFYSI3wXl1cDak3BTc1eYoOczkzgswBM8Mi32pJTLMOULHUVUsv+XjCpMzGc8ckKAZdziNFiujDv06h7XBC+aUemyrOhOgLcxh45RQIgOOgW5F2pn0E9vVNmi2Kq0L75MKIpR+XfdNVNWusz4t7a8EX5GABWUjvJPNM1mJhS0XJdFTHmDwAp+lspdzdC4dQcAUMsUzvh9jxBrhsYeQJffQXSA06pTIM8i6mZnD+TDSz45nsLX5YhN61wkds1qrKNDTid73HggmdWZcdo2vG2gP38mqPcHq3xehz5I2SyQnvNFgGOVH6L36AhM/KrkijRtlJa43XsIfDqLTutRvylWVXOMpNEli2JeRb49Q3CO48qSsl1N+UNi+ky/geg9vu/2SYkULL47aM4rkTZ/ZJG0UURVq8Z6vlGUt0jjyoo+GB/fI0Yuua0IMcxHLF5MumGIc0ESDL4li7/FTUnKsuJoqLas2YeTMaTcuh97QqtjO7FiP8h/deFepVx9vkSutb5qv5gGeMDFqUI5Qw8m9g6SKb8wvu8y7fx3Qh1FuETkjdmaG+QrjoigMz6YYoK6SN8BIV2lTVd4W+DTaLN/splRRwDKVJoJgjIberHkTA114xGeOKRYN/yNApWbZQyt3zNt6xOP5Z8e570CKrEnXm7NRd0GEagQVFP+mGxS+SbXUp060dwSnMgxCNUhBZfMnI+EDcuwwb7Fz9QmDM0eyj7B9RJSx33LkwGo2jQj0nG+DpOe6W844zhz/olIImBZiH2/fd/HY+xNnzxJY+1DSzDqgDrTlwKkIxZ1qCJFwHmn334lzlN0OgF5MQbQieWWGreFwpAwCcEsTASIJszYew641LEh2vDM+96uneK0ogmqup3nf74qrFclgpPw/EsDijuOCjulDj7yI+QlNuR/LkMYgMFzGkcTe8oDIyhNLCRM0HKtrTyK4AviuujBDt6a+PFneV8ezVi1VLUyvX9aAJeHYp0BRIRXmF1KygTK0iuZbrU9lTU9ww/vKxdz3hUnC79UhECXMAAAHl1aNUOhsepfNgADbXJTSqNfaA8fUACHQY+FUVdiPsXCASN/mspL56RQB6tCufdRBBs7MZQI/IgtStIDBQ91e9aaThIOVY/karocD7bjRh01jl3xMv+CABEgShhyZv14zMJSI6i0hqmqnhPmYfIzLBbT7eLZClbzfVnATssIAHQZb6tp0OgZSktglqQ0W0bGg6pu9mi+DQp7ywjPeumJlF18a5mH7gZ2lthtkAwUquuDCe8SAOLESchc10mgcmR2SQdqipv9+Nul44pzLuNIhGxORRsxCBgb+6WfAQQ+f9GgjTxZc/Snj0iwAXPOB3+dk31SlAsi6kkbx/5bTQmX95sU8sN9DRyUVB1lchDTjTWP/87kaKq67LIt1pdVOh4jUuCyIbYLSbw1F8iWhW0uLfqj0lY+qyHlgZeY33UbO4o3Y7N2376JW+HZnOi9zkWYmQy9x3xwTIBPRwjdweWvj9GDgtzxLBXyCMs/MrbU0o+/l/gw1Ymm0UQwk5a8UcgcDxepP1FDUfZPhAZmi64kNBOk1n/8OILRf73Csd1bYP8ZtN0GofdQMGHHn+8L2MlEvRcU3GaQQKLxo2j0g+DyE4icRXEpPeWzz4h1GrsNqaREgBV26pOLsRlPjiwuQIHicZp4rARmYxN/riaOJjUlu+3DsGVwJ+5YX39/tgHh6Oc1kCRGGUGgiIF79feo8RYAfOMB7qb1uJPGKsbKgNKHi8cBisfs8ZVePzt1QDD457tt+V2BLah2NgqaCmpxiKcJQ5kY0YXw3FHbovcUqFB0u3YFODVr4cQkAkYXB+aOhxKE2mRiUZYy0MPfAnFdaB9SMmV5MOxBZ7UazhwzN4hs3V0GRC9pvaFVNvoFh1ULCxzSek4uWItKf6aUKePQwewB66+dwJ3bTaRawYqH/mA0ooiBwgxAD+wlYdgxQwvLORR0OO/ul7XKmRyB85C74iC6o810vusRUHa6x5Di647l4VKB7XHSfQmkGenZHQC0ru/tz1HRnlQmGUqPA8LF6b0rJH4ST41mG9Nlz5x1MdRmxOxUSZUW5fUAvRUPRxJr0Q2l/1vluJPByLDXziUnkMEJphsOSorwjHpzED/Yts6+xXsoodefIjeMs20lN/s5lqc002r3uhpjLDXV0Sm6w4X4lIWShdpWfWYxE/VReg3n/LFxuXrtHy8JW0lpjmPvZqkAomppzO2T+wVUdNUM0CpVqErfrypdA6sEj+NbNAs5RyWQg2yBPRk7Ry9jwuj7T5CgJqCLxWj6IrNc1hr+yvT5DmKp5/TWI653wVcrYG+tahsrDLQtRSQzL4LthAUE4n+d0aCt/vfp8Bpl8yxU3VdYRKdeVrxes1hUIR/ouE2Mzw61F+MUddYCP/jtRPppbOdjjFvqbH8xVNJGHXjf1nn5EM3/WC24i/mEZiKP+8UeuR451ys0rCZnLsrR6YC4jiZ5gwg00pQp4UCUiXyExcmdjEpPkgpMNR4PrpPtYJZhu2QDO1JA3mtX0EkpBYI/qp9yD0scwaphsXOGaqEGVGUiU+itRAlR1Xm3Vn8Q3wsD05YWSdRNxa/HEb87Wje1i8cJVxBrUltUc6fSNQ8wtydDgUnKp9cP3KVc7XXFjYrl/ywCE6DkdSIhotF08INEkLbUw06uWKIfNxTEiCTchON2SOE2yV0CDnUZAhDNt8WUAa6Ae8hWbQpCOGWjd6IERpGY8nbezFMlkBMQz46jydRhPrg5X7+7puS4AGxWRRzm37wVVhnrYiQr8cZDg5FWW4ouFM9CezTAJyW13GSDH6Blsp+HWF4gpCBnfYAo/SZVSHB7l4cbq7iR3+hLdUPctRHm/48L96UaUQVdpilC/VtLqsQa+v+wS3DGO1r1VTliUoXCYk6kPQY0otzl1Z25+7YUgbplCUE9GEzdXso05InH/S77m4Xo6D39ZAkz4TbAGfZaoQEw/mvJ84FOSRGrRPkoQqCeO8rygddWaMTJK1JUpqriE1+PTBj/3faEkt1GBdf6nxKI/3VRMf92elSTTB4zhS3VxtBYbPUyUnHxWxT0ELVfFlfqklynW24TWNKfFfr6bzNi3MPuH0AWfxgLqleelDgjbi+gHkE5d0uV4hXOXGhwgiliqUbxoHgDullkVEgGewuhXBx9ZVwpOocgVFyCRV6Hwr6MjEYPDkEhQvLHWgOnw57l+uW2SfoICffgfdGlmyNSxu6E9t+Yd6mszM0AcRdywvlNcW540OftFkkessE8Zh77MSTH6TU9vZ7qLw4v+hLkICej3jIORJ5wVulFbHRMrsOvYMgnhNfXUjfvvwMkdVFLFvVn6tT4KpB7T+3Zn0zGozu8tTBKe/jqzh099B0v/ICoxjTkQMLmAlj+zvwvs9ZTLJcMhaCk8u6WdRleECTygPhSkTfT6BdLOElahc8hDimhE+CBEVZbzSQk7g2FMG6f3RiKXagbaafUwdlmL+VH3825X0PrhEJPM7wICYzLtFqgOfpRuQfVCdBEJ4mpqhB13xMuDmGKylh30u63hCZ4jrGDoWpV5JYbZH884Xh8XHcmqidonuRSe7fTZXvKCN6sOD/4WvDBP2/Rkc+UMPnHWhGUwTo9cb1CLbp4vOIWP7B8fEtarxEPaCC6jtj0xfxn5pKnuXO7Ix2bknvG2siNE2LuYWtdZPfmoCJ49TgBynKE/Jt4/MTVUWh+W1Gb3lMkXSpJHdLYtAMkmY6G9UYBNhAYCe092zNFDoDWD5H2iyarza6gGF6Xi7KilOcumCL7JkotVlrkDoZ0BYAW5EsNiwzzTy1TSmjox+KcGxA7lm+p1cvk+y68ppjxkDDYUEc/IkqBSm1DUuW1ldd3psJ18H4fH9NcS/snDcLxgGx2iZbaudl2mLOYrY7QLbSfc8d0p43ncqS+dL/n72zby7InyuejcNA7zuV8r0Z/mpNncmwmsyGa1XkEgrx6kWI3u+8AhUvjQaUT00raEkEjtoTOCUkrEkeiDoOzkJqrKp0JxsWHBfXUx19up5UEM2/1H2LgvGeF4XcLiXZipO9FQxx1lztQFiV+jIy+r9z6qJ4hARNJ8ED+xQ+tLKWxX6YK59YxQ1xLe8B7S0WVYCvC6/Uv7zo/RMEn+e+i8VFNYVonNWKLjzVJgyLG9bbF954G5YNMG+cPONzeEvJb30TJA07HBweaFdUedRcCsmktUf9qE6OrgDnoTb62PlGvkQE22mi1xs696exbmZa+ppNg0zh7Z1/6XVVu4+ps8tQV6Krzj82vXT2xVp9r+JbpCyc3wblXwxLvwUd33ARR2fT5THUmMuDjEBMOoXY3QOw0vgS6TdAkAw0TCorty13t+Bx0MF2vLl0+CpR912zGNqWjYqewOX9bYVcmw7r4SBZAl3dk/R4OIOoReI780O1vEXdi7sV3Lhe3VibYTtJFPjov+3H4zCc98AHu1imDDa8NqbayecJGcNf45lRZKRIbLB7hmGrE19t4jNEQrlAp+xXbETP4WdeWrB1UMwPDlI3E0MoUSc2wOA0mrgwp1HidajeNgmuSZAf1eD3QQTEJ4P3EF8KoxLgd1v8laVUypY8Ze3KlND4KLvWtYvZJR8JgOmzqG3v7qA1q1JBg+zhD6y0cb3+/8HUE/q4nMWIlkWm1aa5oxsxvNmPp+bjznvRwW8GImdAMWL0ThmmWHqKF+pqe0zevKgzH9rENFgbPCnNk8PWNRzSa+0OtpRrvlFDHuhbDiQE6bKhUwmD7fMxmZnAqcHPRiMqzhergvn6+J+qMjrVY3nV65807ULn+gCJfvnwQS5pwzKFVyUbQKjschDsv1rGajRqlJ1fFK/iCWrhvbUI6dF/o4nW5d81i+wcg4aQPqTskQ8r/XMn1kVuTbj4Ir3AFg0ZQBNM/IVfm8xd5O2Jf88pByGX6B0YYMPfaIYdPYVYZpmee7fxaTIJNNq361hhvoGTlR6FVyqyQpSUTOtZRvrucoLQ/WTBcmKwKWu+FjTZ1MrJrlkmTfXaGYHQ4dtxi8R31H1MFSWV7fHpsEo2lXhxzzo7v67GWeplvtERoGo77NyF1zHztT7oAbzmpvzf+P0wEh/F8+sKnzTOQAAUL0FZEHyuIKPnmKarIk1nzRa42cNDYAATCxNZsB7gAAN0PxZjx0JEAd2eyLcVdQVjPPQCQs2ynlAJ+HSE7J+AhaBfFvVBlbrBkaIDx+aoyOUEH7/YSpETwd9XN5ZGoYKL1v/UhTTcrArNg+4vUDrTAAS0LLICiB+M6HhwE8kdf7P1v+REAfCJ8VeLZyj7tGneQMXSqdIP4SVuiWBHFNxDetJGBHT3hrbn/zbQuhvv7Amx0v6gOgnSJxavChUxDkcvRUtk2CjjWxosyfYrz3vrsbAwNqLD7W2U2uWiqeQq1BTejw9SE0aI2aEC/ZqDyRgQx7UQk3jFw7BiGjokunVdPWkaDolet214fve3JSSsoJQbiYzeCb/P9QoJI8bp4sPPdcPK9xD7oHTPvdtp3dEHLhDMCGk65EHJYk0fqz8on31iiTP2yYtQ/CdMQLmuuu+FjnYEPvFQDJf/M4LGC/u/CjUBpuc0sd05X5YUj4OhwSzWIDrzFkgHFT6FbgwXBdjYg5sLciv9LVlYI9/cipWS2CbEtH2sTlsIzRGUOpbzgegOMCUfS3S72htNftniafD9PZ4+XcQAU8ZlBhkQvKAnM8PdEgYDdBKkKL5yPfIxYbm9MRxbxwj/0eiyGVo5Wato/DDeTJEqkFLJAGhzklHq5eALp8AdblEn9AJfdnwtKt8EztMAHXVsDfe5eBy2WUFg6NGVIjVxN9J8Phy47KBkk3a1qmUeH6tNkZDjHiNKnFEOKSpE1oFrwBr8E4SL9ZuzDySV1kLYWMbrI21aMMAPxxpv2AZYIWeoFmLVjzz6acNKYkTQuGd0FYZTh6toBHZuWkFTCACYZRFNn3kihAC8eLbz35pIvj2+MA+uvt3pFtzob+wot3nFI27cVOUqfmr6u+BYYpwnHhvBYzZACzKPa9ZG5I7qsPgGKVAGKdNYHXOmfoX9A7wVNqfDXjiKdCgZUN/AoZb5HKElX96Z3YfhrvEfucLO+rP32XsElcP+YL42kehR9ITKHuwA0GQrNJw2/mCxLp5geT4XKCj7HOYE1RwEzhs+zhhbs7te22IK/vIm8ct3fwf3sHPg6MvE6/ndRG1uBGB8ecYSRrcfAvio/EA/GgbT4MT7pVDuM/ER7TRK66TXu6sE0kqzm1PtEYMaMYHXpaevZ1y+p6youBui5VO5ufG3a+tX6Nwge1JYuuiG86PpcLij6/eq1vDVEWSJlr8ltqxlkIHI93evZtjx7z7ToqzaDwaSfzUCaZfcUDXOpus+dYENAQGhvBone8AyRGxbqwNaYZz01qa7yF5YdwsEpoggDZdjUVlpDY5HRMoedM1NXyNP+1YrX7MtVPoo1GiFOOEzMq6iTOQKcsOFMOjCOWBcn0qKcZvWeT9YRBBUi3j62l/7fPPoj3PEf8t9Jr0XBmKISEimmxi/+4188akGmEwCR5+SJnCHJE/p9B6TowS6ZmF3KF+CacJGqNOyHttu13PIKCUH36jLNjWvzrAu6cLlGOASY6mJKcCJymAI1a7gOXeA6nUn64sVuqad/xJ50MC6nbCNy3wgruAKAGUugvCN/XqVfAKvL+rSlMS5JD5tvfAyC7HTS9c3Qkc08Y0/xjr+hzcM7sPJJ0PdDql+zV0xAd2ZCjVx38y3GpRZsGD15hg/9EtDQm7jn1dZAv7wXooRdKUlP+OCw12S0UbDmf18vYI/uDXpRqWIGJaSai9i1yv0DB0eECi+SUQDa7QBjxxzyWRLlJ+dQWgsD1/cJQ2GM3qGjHFfVV28ObhtM5Imd7wD9e1VPnLLit4F+dCS5qwDD9qMrg1PYMMbOGaipCHoaNmryRCPoH28f3EZpn+5B9wc7iWLEg3ARa62pv7FhvTAPIM3Np1BkL52mkiKkOFsrVBBheKYCtsyMvP+SLiZiiZ16yp7+paMQYZwTGG7gwHYGaLkbyFWGvEHA2jerHv6nfrr7dFJ08qNkmj4+fm/qEzkWp6wHw1a6URVB98I1z5fF03eVyFll8WXi4QxP4tZqhET6MPp/VBlEWCqVTgG8HIljSR2j3KlP4iFfKjQNzO4PI+dQ6hZ/WZ0lsVeBMddz0wwiYDKPo42uaBpzXqCYHAGi7i0Ct5orh5h0GkNcWiSrS2aNmILW9TIb/qj32R/SUw1oEjaX3Tq2lxqtijC+wdVTkK5wkCMWAAkmAvmQkdoKjRpJuvzRNwjq64qJuaTdK+TIuTNXzOBHwKlSbYFVqAC2Zr1yXqbkWfjevjdGSy0edDmctZ6vc4CkBkwHCKr8QlAtwwsCKnsAAkaUe15F9BeMEDMAAAMGm6ThIkTQs13o4HLDSnvzVgOAn7r/O0xxnGXWSGIDQlvo14ROKODsvxZbqNn5Z8HqWvfk+T/ASc7FpWLYWvt6+sgInt+Y6Qz/boAx07SWWPicMcwwT5UvlpCOdF/H61CxTzpmnUrWF/LLqnh/DwMFGpb2CkGJyVZg96v0IXzjPIorUFRU7ihZ9WO9QjUvMBn7SuM1TAairw5LCGF2miAAAQeQBsmW14dC5v9i7twcos6dmWyoG4U7TBvLxzXtcldX/r10gzkeEi65AfgVO3zNr3h/Nl7RKJGW9bQaW+AAAA63bt9GxAwQEi0rYEHeC+80Uaf+fRe81cDgFiSHe/cM+0mOZ84MzGr5OMsFnYOD6Is4OjnMcdrZpTYKAMc/b5KDiFNH4O0acrf0cCbDBqp989yi5nkdumFz56A00YXajrxOu0LEt53cJAIfHsieaKCbTntSnJBv7Ob6v5nyOB1mVJuw1bjLMjVVsTofoxZ2F37Wp0UkO15XDQPgy5MgHHBoBN6woe6PfhIwZcJ4NOpBhRiVxSHeNbdss0zZG0zplT9OoCFdDyi9kCLbegYGbN454m6jKvAAAARBSOddf8GL4amS6D3+jQReDee7uTgN6DDEtf99KTqL9/TeFC/+6gcaaiUWK88XOOgyg5fsddUXSnVpg6FgcfXbuoxTEKmZmRhD9QSjywEl7MaE/of5gGQNfBy1DR8sR+t1pguWjr8boAAMExgq/2DyVloqOxUsjyPI8LS9bBIkuwJO6EuHwT5QveoDusyJPmQlEju8dnEvQUTwAAAAA7AeGcAn3ADe+gAB3LUgewxAAAAAABZAAAAAAAAAAAAAAAAAAAAAAAAD5n9alQlaYnRSir4sAAAAA1VFuH8eXAAAAAAKJoZ2sVEQ1ThZgYV3ghJmjJn9sdBzuV6y388/IDXdjgPllv9PrRIRAAAAB6pvPnuRdNk0D602/v4ifNouT23b8p12c5SakvBJ71CXAuIp4P/siAMBA1piCQsUzGpmQs9/T1Q9MVCYvb3LA2PkMZ8AXPFcENwh2wK+8T5FPA1iI5r9uzZLBRdQgZx8CtL+odfbenPuGafawKBQ2oM4FrsLixUnzdKQvr2gthxG/MxX3j4zDrVSqeX4yFaQ7OqdTtzF9uXwAEtuCexVMYn/+K0fVO+yCmdpVM6KUZks5d2kK0VXfYlzj3eFh7zBZjL5MDgkg+Wg69fs+PRegclPyf4sBBHKWgl+UKIFugPYdDKqxes+ZiypH0qs2YQcHDNupzYgto2E8Nm8/QaIgj4YqaEbXnkeaCWkHHLVtVw99cVkwJzwfmPtQISi9GDxUxL8Z8Pm5szZtM3hCtbRIomjGQU9AfTp3ZMTAQcMoG7P9o1TdEa8xxaNxlveotKwGEXFmvi84oVQNiKCP61m9QnI+9I+gnkm8+469wZFCUAh/CvnO2BwyZBanJ85BWiDf7bD7HNY8HkPnQYROwWLDnn5U8DW/J59zU0oFA97uwO07ppmsKHI1lesKhy7tuySRuG3tNFtgYgoVESMsq188KJto8R1hKQkJmsbted4EFAla8AEtEast4sxU2+WjzMc7iKFxvXxCnK92KMeK/ZTh1jFFtqHfMyx4mSzGlIdVVCL71xT73Qk2EZyp3AAAAASKhxzPU0sw1IfUcsIhbO/+Md8oPigJ0ThIkw7a6kdMOqQAAAAC1kWakLKZ/SWzRy7w9+JuLyxUk/IslZ1B7MFYRc3YiAzagtt6p8FmwmWX+bv4J+kK6e02i+dcq0OBZVqWddSE2neKIRrSx8lsqzIgy+0M4Hlt+EP8T8sY+m61z4lF+GZF2CBOy/nNKJBLxzRWqsA6SakfdLtde9MXlLbvwSeKnZfI5JyQS/JGg1lPWFfapFxiigpel6DoU9WGGZe6yAABFvzkLvmcCxHgbYD4aJPLPq4fdKGn+p98zAHe/eG9VKW9IQkKj00jq7jtldih9d4a/ghW96JFJNzIMLQFCr3I5tLqq2fe2/oMJoIFchtfHMO6rtqeHoOyqMYAhrW0YvkvPk5NsIpaykTEdrpJeXw44R5eGn9cSf8ugCD4xakZwmxhte8rnb5DktNfJ/PFKhrlLOGC8un1QGYLS6TEpNclnDiD94uf2E+UutIF5qTVhTjNPapmjUPtgdREDqAGuTBjtpwplFVeKOqNbq1Fw4gdvxot5s9T2YMttnoZO3iz0WlzUD9MC86Mhp0U95NaU/DM9MfvG7Nss1X8QTAT/x9/RXKrST7KGnydzwlc2kTocZ1jaxQa8aM1icldH8MJqHW/zJAJB/lYeeNx46L7wqfupJv9K/+ghjIu6RkdlgTAH3nV+YS7cCF9/a+5Zzog7InQWNowTub910I7xytEB/g41qgTTzIN7OYoryo+POGtqXG06ZAF2UZ1oc2oQqGgH6i5QzJ+VOb3dzvyfl0m7a+muL0a7vr05SU9IlQscwhbDpohqtQi4vdnAQ/zZmNvBYteJB4ODPg45MwtPTu9qLK7exrrovLeBEQu3oRrnIlDYf7pSnTV+7Mgk3J0bkRDaMlws/FN2ZpVGxReH9U/5UOBoqLSnE8qVWUXNjiRiHs9vnq/PIZ6dmLuFnVU2EVtQlbO3g0cV7dFVyFObobFVtNkxUzfuPvXzWy+iDuNUoJoKDGqSlmwtvqpEEoo58WAWS+wnu8nCu4Q1tb92yDxKvBog90/eVeT8Dya9PMwMU2vxMNkIzupT+o1F3w1jATn5c31FFiy0fJ6sI6+6L4neUUKn+/6ymnF7cAWyuX/jxGBsTfOjZEcO3pRHvG+xci2G+qRs0UUFi/Z8kg53rnrFUPcOJxN9IuEWL9Gs2Ei34ffEo/q8WESDRQy0XaaJpnMgK4UX3mSNUoVX7TXndfqycEh8a8ovAo5lQexRP97x3y1uu21ZkUhlHzhVXQ3YdqcYrCKr1IkwDdaQx4eDkCza1g11UYctjzh90Oh/6Xl9APJLeyOQ6Whuz3xVFVM39oYPMaJfUBUQddEefO4lmwhh2ONv+U/bP6lJd13zuh2AQzglKvOASkBID0p8yTNPWLjYy3V5db5yRj5+GXpdZ/pfe2ODpD6X/8kDMhWbkZowfaYPV6hVj+sp64lu6ycw/zkHd9F0oILaDGF0zGCeaV1aI6wdWJaQpmxC32MnOCZk9pigLf1alCvUWTBbR2hcPSb7HOOxilmpiBB+J6hgDAi9W2A520+YkkbUH8+NgozY1T3kn/0ngBf7TUjQceU4HiUIKW+6MXhE65upkcwO46pmjWmi3yW58cA+apQSrMLRa4j9SrGwPIgQkurjd3gw47nIaINToOcksJ+ZEY9SmPyghGbJAeChAs4YU0Axih3jm7ZuyEW5snqyWcR7lg/5lw2X+MRJAKffJWF26g7fglKhBN3Erw8xEY57eoxbuZmbWMVHescm8dv6T70b41+SLYtWbUzDiB1RIfWcr8aN1u3he4c8aNiczDVQL3h9n0fKA8/LuJ3OJ1cFbb5XKp6KDsplbUJ4IO2wjgx34VJ9NTURQVqwXZnOU8IYCCDqR4i5nb2640ocmKRkkNYZhbTQTa5mW++PUh9Ia+0upuV9ByVgfKxJCVf0SVHljae1hxuQ8gegh5F3u+fd/BERbO/UZl2EpZztfGnpCHg/5/I2TBgyLOUO6vzjAOLpEywGM9qHAbbsN4WA8PZMdgHM906E6vIiDRVzYyrgehRdw3EWWzR1Xk5yg7HBokIaISl6aXyrFj5DXwx/yd7NdfPc3omIa0dMxPtN3fRTTOpVQK4G6uecaGSBHb2Z/dz1Z+6ofFbTDz1iaNTbQD3Dq66irzK8CRjPilT3Vj3uDZrcNWStO/FoH8p9W4jIZDoJpRKdx8I1jqpBzSrBZZ+yne/B0Xs+p+M2CNf+GMY127YTKAF8xjxGh8CXKKwpBFcaGTzLFD247vqMWZxZu0zAD9RQ1tFwMxdO5lro1yaoRrCnsshUHECI0p9a2BUJT/wPpGk6GJOFI6M/Rd5Ly8u7PFQLWeimEY2uORc9pDkNhp2Gp1hGN1q0lUp9m99/a440Gpqjo2ek/vqijMskVkPhqtjBcNJDw2yDK7dCkcfTkh9R+j4HlhQSjfLCp/Ef10GmYLTTm1uEn3f+BlDda0DcWcG2DOV9FxZwQXddQ2i+fwg0Dc/MAymx8Rj8tby7gMUT8Syrk2NfqfOADCOBQjH78TOZs1llLzl21idyR73Wlqe55qmwj5cf3j41bs0wdFBA50tNqbqnwBehMfLW/hPvHmsSeZ9cldMCoXpFYqhE/jF4MW8OjEyOYVQjehTwiJAvRvW5Y2L07WCOV3dU4IgX54UUhi9OmK6dvMRp43oOrz0efNEw/WDS1aMls2+QOwULs64Lm1TmjSzw/V9+2E7RrfZZeSzrZYuBdM2EOiQZM6yfQWKoFRmzUmd0o6uYE6g6wvMyns3EEdjnlElEy0qRsPzKTtZrXzMiGJyHPZUfAgaWGhtni67ejU0PTEoQOtLK8R1ld/vfdjDvXP282eDDdbjNehXHq0cN96/3KfJQXNfbOpDJYR+IwGJSOB3O7njyMGYb4UV8dT4PV7Us3td/tzU/eF++lr+4sdt7GZlF3rughIWLhFC6J+oI90T/xHtRH+m8A6Dm6twfkNDiDi4rAOevpfpSS9N+eVJ+muDJ33K7PdzXG1kbcIH4VcuZADjM0GoZTL3ia8Z8b986TzEgQilAs1eyPDi93eA55Sbgh8ZlgMsDqXWCG2aUUH5I1ZcacP5zTLSBFjHbCpix9Yhhtl+qijuztAgvdZy/Q/wtELc8kqXJaq8SaCsLag0Y3+0vwN7r0jMNP4QGH2Zb1jr+xbLxHQQho6PRArE5jSJ0bF8YVJwgSYpzQiyVDzT8eji0V4zr9varl8LAY7r2F5t9vaNtTm2FEWzjAsuwZTY0S9JXOJmxMlVpkwfCR2lZYjAmWtXsP4uo7HdLa8KQUqSgFp5ZpFIxTE30TDsPg4gjHOnsArgbfheShTI7PsB02OS3Lci5eir0gMRH9WqOzLAewTpi56hUT6mBrepsYxWHWAbwgLofnZjq/Y8y1LibdBJx7rsRgMn0PKIyQ6NqQRsOMptpf8x8Fm0R/mmmsmT/0ODnZpqtZK25RMSFOLA6GjgNhltgJ4TyYlo8q+mxUuoAe+cC31TzrXIFIvG/Z7JbC0vzdyTJYma/E19Zfin4nX/3dpyIACtSkTHCkIDUkqconKqtiaovIghzLP6wPlZARMZk2jfdixvm8dYHb3dVJv99rlKYVAmQrl6+/gVakMhv2Wn+WbdsQFN+vnsDzKq0yYqDVKh9kLc/4RHFPxWnurtavTd7Zg3QXCDrHGijhy/CWg5bz2cSxl5pCtJ52ibvdqdD0QtFQ3zjJEyA9nBEjJ0xzek//qgInszuemanDlEl31HmOoJcCzd+12UGTwE8Mo2FLbrkdEkUkpO6mSGLbvZ8BcnMP4z6hv1VLLE4UwMcxFtoHf6fwA3iH8fPCl6MYzoJHp4FFri7yAEl2+R/WRWfuS0Ad7WXbZBoNeBEwaobiVTcDzVVuCp8w4IV6k+0Aki83MugxOxkEtFXfCU5u+d3s0NgoVE2mPsDXnl7js99tDJch8RBzW0S9m1umLRqzQSUCXnOVGUqswhZ46VyQIJgxNTH0LpACct8qgQdnKNi/3fekvboGStgDp0BidBfeQuZZ/BcEFmY02yA1O1/kJszUzEUFc+RqSd3UnyxBzKWoMhkM71L/WZaCVAksLyKQn0GN4Fy7PxwooP7510JNSIVWbxVWY6x1500BQ/qNLgSqXIqskoA0OqDTBkiWbOfg5Saam52i6qSG8vmmyGivx951ZpRrjBhA9EbNyChkizJsvOZDOso/pE5cI5qtdHDC8lSB+BQVF4dtF9Oe6kM0c13rg+0dji/EMco1WAR/RhzbnPv4fi2NL7MYQLXz4Cq+4V3hJQDRDTGBoO2SQpDDxSiZyqldOiGfzCkeBBj4TdJ8nLIs4BnUwB/ts4ClQhhUlbLN8aWdqdqutAg6OE7AHvq8iOMbn4znm/XXK7MmIG/DpxmorNVmLB5jSzkAEcGULVPCj3XzSDg8IHx6uBrKwhJanICHWTUlGa+0reug4L2QeWO7LrPan7WTlFqBI/YlyB8o4qMmLG9LiECNGn0jGv8L/X2F9WGAMm/nQJf+Mb64Nf4w9PP1ZCuxI6DkKX8ILn/ZX7pkT8ByBG3oJXsKFuj1gwTIU+uiVqJiAkVyl2S+fWjiejmEWRQAdzJl/RKSVD/G5OtFnpQnRMXYmXq6l5PjDR3jw+SvlG/N5UfUnPq4F8MkSLOpauUJVM6qEVFNSRgyFyDIADJ7xOlG66LwHx2ek5gS6rJHhOT006bE4GkazbugAyaizP66bPlg5FjIqqDBraO4RSDJLPvFQUZZffkDaj2X0XaCnA6vomb5xdasmuvPZ6voAjGjh8rqbEh6JNBwbas5niOGMV+Jgo6ay5SZT8SLUOa/v4co0fj+SJdT347UWIHj8NbtVRcbjMtH3xYpJORvptXVgn1sOzD3ODi9He527kAFxLW7lbB0pBposznbeyvWuU68p49UBgXSxR/Pzu1xbaThb954wd63uhwv/sSmXJ6wg3FxnUV0eiHyShMOkVKdJ+qstjGKSV8v2lOaphqYNsFdR4nk/XkGZDxprw4CWJJ4WVE/SiINmiND2ebPsPA8pcAhklxCRY5FK5z7V+hBp3LwDB1gzjT5pTRTs/F1S4hLvhgipZSo+UVMs3l8S+YYVGVMgIgDrOdbpjHTjw5bHB9y6bRfbQu8X9NCbTAIr9/+Es+caHUapBz0Am4H5zShDqBOxsfNMdhcxHNFWL+cXh7Uz3UbsRtIfb5Tlemo6kmCLRHHKeagGhfaiPnTxEDYhxt5v6rZu1DXXxs7aYgH2e5UOt0H/vgkGDcc3/2O7U7Aq9Qa8ccSnnHI8j4dkzwZozb6ko5yj2oN6SnfnR35kvn0gc3vkX5q+Ib7cmw2PauuCyxAUdwkBqTpa1M2plCgDb+gwkucxs66LpIamp8DW9tnxZOpqf2bGT6JtmQWAuTJJfuRH4f5T2KNSauJlE4l9LnVhyn7Cz8kuT1HMABd/HIMo3SxR6JyZSZapdZCndzC0BFXisH26MUxbA0qyFTTWLikj9wlLTozEbInQiYD+2H6OygAAbxg5AAwVX++TZX80D9O3Gy6Ykqu8VKcMPkrk9mpbVUubvXF71gbN0x0DA9VKBE1yrG7KmTeay45lXrIRTC+uZGcQ2GT1ny0lJa179qlgEf5eidp+C0NNniOJYlFMP8clcbvsNp4tr1GI1poro0zvBfF8RokEoIcerZsiSGO/avLMMLpHvZ29Y32WRaVqfGc5pf0VEac5AoORJIkp+MGWa7fQzgvq8Dun7QsgD9hgdBmhretUv9YEwZ9ItVEpT2OipnOmp3Vby2i6n2R2dPFFeIoT/FzsF+X/9/Dj5XXGzKIXpq+BIUF+qPyHn/yDqqHCTJN1EoXD0IREwPfpWBbx8u4cI+CVVaorGrZD0EmHde14qK46W3+8tJOnFClkMYDe5Fy00YY9hcovYxFMG9mmhlBYYDiuxgmj9aIvLsogF68BYvZK8iEOQkH0ZsFiUisHKCRavT/3i4a+y/geCSRaS7+bYaQFUNTNhlrM128KOte4LN32zKVjlKfYQFjkp9Z5voAAC3rIVomwz/QNPY8GZlZ5CY4iDtGdCWJkAXQD+ALQ3/saAB56ACFiAKnkmxa1g9KPeqAfR5zwRxuUGjPYzKk1wJQD9mFav9txz7t9P5nd1fXPk6UbX3q4wR9MdB+1HDZIpwqlU8W7fIapWSRJPAlW3Va7JnY6NVTL6ScrgUFw8P+tANWwiHoJFblnnoznwwWUhLRMdOngabsB/fJJA3JvXhKTsdM2pZSU9C2VFJuY7U709djarBGGCuk3lJki1Wi44buEdURlsUrD2YN/RLSk3evLShtZgm1af6yU2Y9GaRe6C5y452hTCW0eU0ODyxCG/X9sIvPJqf10D+eL9Eh7HtIa3p+g8/hQbl/amug5J4d9mwKmlXxCJJ3aOg0qxWPBHHzXNgojZrFqIl8d32kXhF5rI4qn0nrJ97Wujf/NeSd66mv0Si9omoeBVduun7CUbVtRFFZBfJ9oUNYofsnxtrSN9IMi9cD5jPnoGoU3dzyBlz3urfyWd0//hORfLAyqGZTGZTdQro7+eF3rVTaXn+Fb9s3jJgpz8XThU+vFd/ItIe59qFwdkCenAdFkzO1oeK59mBPDKoU5cbo3gQVztnUsoYPa3rK2khLBuq+BjA4Kam33JTuYiNgCC5pesEAt/i96huYPfC+l16eH1bAv/raIhtTnZJtLzgzPW7y8sZ5Tf5MvgmGGunE/AJk4rHAfSfVBycsUG4shmYbBluw2i31GPz5FT/4/bO+EpnE3tl4gs/eZrklXO7Bl24pM1goEcYkznTtYorxPRx0ZP+3uvycASud8nQmCgIlyF6WYEeXXJBwcrceD0Z+V4Wl3doFoweuA5BXmtdlP+S6Q69j10sq4sovgOCPsFVojxNSt46d/xo6mD4BW29I6mS1A4R0466vBPPDw07hmgfSUCZOGdsynyu/2Yc+qnQv/BLrttnG21rvbcXh1/EDVybHlfjOgF61LYLTFzlSUsHYO4yRRVEMxSG5Z9q0yFq6cJSva/uQcYsc4u7/u2QbN+9aXFF/JyGY1R6X8rTIVIz+aum+tTsueZn8iDt1VVHESitR0/I/fVuyTy65ymm1DjX1UYmczSZAnTQYyRfo51sSxdppFo3ugIdwGlysR1K/AMnPloC6iMwxEXjVsWuDfuMSnnmg4HraioSwrhfBAKDgw9Wk5+7Jz0aLnNlAxU9jqxhcNEiwu3/vH529SazIw6UiQQFwPXKFj+loAru3Z7APXAfKAvV/1Vvuau8/QVaAdaeItVsAh2pb2ufXmI+yElGsEFKAyJCXjzEMDlMft6y3UqJ4tXGG5spmpQSU3I4to/gbnixMY7iO+3mtHsTCgIaX9zdm9mQQliDed8o7T5YWHNrDaXfgVvlShOGJ8u9+WYyeDkUfHs2zgNNZEeRH1d5ecYcFDGrCd5llROfW4lXagKwQLqTmJ35kSrc2CShv9lSP25O58PNwm6F58t6WefEsctBsCYiAAgaR8CbE3r/qQX2f6PYpcieqXrVgaBXuM8amUWCxhrbXM2GNazhJ8pCuk1mEFUP/KOO1V4aBVbv2K3o7IkT3RSAi74nirY9gcDlT1NIdckhEl68FJcawdTPznGwri2qLF7urZCKA3am9RQsRsb7bMmH/7i8jPrmZPRFDjYSOFtJ6omAHmLgX6YwgTbrglzAhGxLQVc8/ZRhfHokgKjPFxb2ssn8+exBeabhkL5Zdcb0o/iy1HjP5Yg5WBw4KWfMLqPSO7SDfJfWWmCrPk6bXoftV+6AVKvE/WlER+tMazpsl8yTIf95ik4EoGGZqNEGoOjlvv+2+JPhFIKNuIr3OlpLtE/WPoy11nC/57GhRJNZz0BDgnVFpN/NTJdgV82ixRIcY8ZIVTuXFq0vurTsV4zBogY+3pTfUt7nEv/CGiv1afQvNazWwzpu8bjasCjBg9Bzxymri0/vkDm1593nq8Wn6TTreGcxZv+aFVTwLFItT13hb7WocaxHghisFpIlCJUJryov4Xll/Z8Nj+Km82KQdMGvtG0KFA+09TaO0z8K0TN1ILPbsx60q2ILK0sQp1aL9tIZNabMkK7ObkIp+Fh2zglzIyQIH/jAdCRWIjiU9cN5HGrP9D6iaWyQBs+82wspGo+ah+CCfEB8NjIvg+WB0bjfzVbRk93YdLHc4an1sVn+Q5SUYf+6lSI9DeVCKm2CfpMZCMFQp6oI89qDKkuO1DbUK06oI9VkGIghk2rKzV8CcitU7TraVftnBiO66aJ2AzI2q9IMowd4pJQ1V7FHcZDLmfoNk8Zav7l3u1B3dZpZ4OGzzR5H6NtpIH3Rq0B1PDJ3sII92DoBPERYL+I+Qw5HtEnbHRbGZ295GNCQ/xzKG+jA0k4jbl2Vl49VMYZZqFv6llFMp6yiddrfMRXVSJVbnrTQ/Pj5GqPjaiqv6Lj7Lf0Olscj7Qfjnwth/oGlF52qs6yQLaRNdCfm6DasFQUS5FnOnjMWE9XJcLuAZ+ZJ5ZT41DgIxabOl0mo0p8llkv0Dr1+KzFtARpe27oTp4j8maa4uqdyAep6vABCilFgn57QhL7pm3o0De3fXoutH+PKiTYIZY2B8WUfRw8eeDrGwbXGkVJLYcM2f42ND27c/uE7hW0UHc97HPMR33pQHfZ2iaeqdq94pWoynZ7yhimsTs4pmq/S4UM+rElnqk8qcwjWL/bWP0nT2bSAhxh1uusDKVmuB5h9evPLHwDQdjFuvb7QY50ca/1k7KwFuPP9ME9zE/6w4HSc5Wqg4vsH9LQbt0Fzlj98q8k/5UIfu3jGpS1H2AVb+9tRJ/IggoGSNDNRIjNDczac+BAI/3ZyT/T6d6/LcB/wkD9Hmi9JRS22LF8xJ0/nFLS6uU+RTcdFaW42fYei+VMmODniBLhg4hT7z8qiCEYjrevYQshohFrHkDIp4SGK0ni7HI8Vk13S8aWMMbY5tvEjfTBhVNFqt60z1X47XSj0dDm7Jt1XGrEYcvr6Rv7SgdceQtyJWFp+sA2kvdFclM5QjQ/NEZ6xzo1jQqZUrk671zWmBkSblvZw5D6GxB/f+ag8gVQUJPIVsk8TofAXacYJ29vlt9p14YVMDAjA4iZ8Gi3q/wavZLjmPU2u53rrs59ZLDF7uP7HIayigBW6pQABiFAAAJ+IAAAQMRavyd1AKJLKsLWDGH5k864xpQorDZ7J6AAAAApHUDSRbtSpu7/efA1BOj5jz2tekLbMiY/kr396sx/jXr2mGSXzlGhI2DCNJ+gEYMOvSQSpXskccjgh8cGmBq6MUuegdU5ebGa1xsVgcA7b0t43JOdrOF0AAAAnNYLglF6ffhwU9IheVJgiY8qvQG6Z9OMMMOlW64rx46fbjqCBRKaUYTuseUk5zLgljlrQmnhTlsDG6OkLw7yhs9bCPta013QUsAw1IeyQFQwLC/5Z9kb0okOgAMIzr8r0GZMMQr3g0vXjNZ1iMW9lbcAqzrno+kqMcLLaLr5tCmmk2hfZf82mry6kYrfA8djU9bHKcNc3CZ1vqMEFH/fqS2WFtY13Qwp4jbuH899scr307Q9vyUjTJU3vZzzbh7k0fVNqJt5S6YFy4vT1Xjw31Z3oYFAl+a1XYBCkhnM+hMUfTNFGBz1W249kUZWmfngqQ850AbJEbdQnCbRjUnTc88MwWxRObWS9FucLciiyOqBNSyD77mEXlDQ6w0dJo575xIiwaTj0JA6NK7qrmESw3sf2b6YGclCBvPpqeogB1vybUxlGqEJnpnfODs7sUgrYm4nmenCrHUys/YKMJEHP4HQ3svUpg7aEDtzc6E6hDXGsCWwFCGwXHUpwSb3s59UsVrAy3C0KRJF5E4An/FQ5zcOQ3A2LUH2/7KBmUK6WUdF/VMb8eUijwdOjgN5KRo6T1yCiFlt8AMwx/AQrlNqEwByNuaK3YNvURXJ2I8ldbmU/jCxSAdlVu9x1a1ZWxReo8vMAQ+TMZmvR9yWfNUVzHZaU5vHl4H3Zc3E62KDWUtBAv4fgHhVEXYh6Ok7GBcd3dtXj2ZRVZr+034ub7MMEaX8RMr9xnbewmJgk2JpQeWnaX1Qk4ogRdfNiVOwyxh2S8RES2RAAAAFEMylEXbrtYVPrMtx5lJqN7OfeSFp63YSnQUGhGwrY/fZGWwAACZNOvXiQmeqEMcPH/T/MMmGvRbTMz9UxHPvcYuEAInraRWu1GGSeW8Cd/yzBJ71x5iAAAAA776zDSGnIyhViXRW2t9tPWAXaU79+4eGau89MzvbXBf/KKBPu+7rcqmcoAAAAADh9qextwWlAAAAAB7GMnGoAAAA) Note 针对 QCS6490 演示了 gst-ai-classification 示例应用程序。 2. 请运行以下命令，以便下载相应的标签文件： wget https://raw.githubusercontent.com/quic/ai-hub-models/refs/heads/main/qai_hub_models/labels/imagenet_labels.txt Copy to clipboard Note 该模型可在 Qualcomm AI Hub 上找到，其对应的标签文件在 QUIC GitHub 上。 3. 要使用安全复制协议 (SCP) 将模型和标签文件复制到设备，运行以下命令： # For SCP, run the following command: ssh root@[ip-addr] mount -o remount,rw /usr cd /etc mkdir labels mkdir media exit Copy to clipboard # Copy files securely scp imagenet_labels.txt root@[ip-addr]:/etc/labels scp inception_v3_quantized.tflite root@[ip-addr]:/etc/models Copy to clipboard 1. 要获取 Qualcomm Linux 开发套件的 IP 地址，请运行以下命令： ifconfig wlan0 Copy to clipboard 2. 当系统提示输入密码时，输入 `oelinux123`。 ### 后续步骤 - [运行 AI/ML 示例应用程序](https://docs.qualcomm.com/bundle/publicresource/topics/80-70020-50/ai-ml-sample-applications.html) #### 使用 gst-ai-classification 示例应用程序运行 LiteRT 模型将示例模型和相应的标签文件复制到设备后，请执行以下操作以运行 LiteRT 模型： 1. 要使用 LiteRT 运行推理，请运行以下命令： ssh root@[ip-addr] Copy to clipboard 1. 要搭建 Wayland Display 环境，请运行以下命令： export XDG_RUNTIME_DIR=/dev/socket/weston && export WAYLAND_DISPLAY=wayland-1 Copy to clipboard 2. 修改 `/etc/configs`文件夹中的 ``config_classification.json`` 文件，如下所示： { "file-path":"/etc/media/video.mp4", "ml-framework": "tflite", "model":"/etc/models/inception_v3_quantized.tflite", "labels": "/etc/labels/imagenet_labels.txt", "constants": "Inceptionv3,q-offsets=<38.0>,q-scales=<0.17039915919303894>;" } Copy to clipboard 1. 将 `video.mp4` 推送到 `/etc/media` 文件夹。视频文件的默认路径为 `/etc/media/video.mp4`，标签路径为 `/etc/labels/classification.labels`，模型为 `/etc/model/inception_v3_quantized.tflite`。 3. 运行分类示例应用程序： gst-ai-classification --config-file=/etc/configs/config_classification.json Copy to clipboard 2. 请使用以下参数，以便使用自定义分类模型和标签文件运行示例应用程序： - `--model` - `--labels` 1. 修改 `/etc/configs` 文件夹中的 `config_classification.json` 文件，如下所示： > > > { > "file-path": "/etc/media/video.mp4", > "model":"/etc/models/custom_model.tflite", > "ml-framework": "tflite", > "labels": "/etc/labels/custom_labels.txt" > } > Copy to clipboard 2. 要运行分类示例应用程序，请运行以下命令： gst-ai-classification --config-file=/etc/configs/config_classification.json Copy to clipboard 3. 要停止示例应用程序，请选择 **CTRL+C**。示例应用在运行时会在所连接的显示屏上显示视频流，并将推理结果叠加到帧上。 ## 使用原生 LiteRT 示例应用程序运行 LiteRT 模型您可以使用名为 label\_image 的示例 LiteRT 应用程序运行 LiteRT 模型。其中，该应用程序是 TensorFlow 存储库的一部分。label\_image 示例应用程序和 LiteRT 库与 Qualcomm Linux 一起交叉编译，并安装在目标设备上。 label\_image 示例程序执行以下操作： 1. 加载分类 LiteRT 模型 2. 使用delegate 对图像进行推理，从而在Qualcomm 硬件上实现模型加速若要使用label\_image示例程序运行模型，可执行以下操作： 1. 下载示例模型、相应标签和示例图像： - [BMP 文件](https://github.com/sourcecode369/tensorflow-1/tree/master/tensorflow/lite/examples/label_image/testdata/) - [MobileNet LiteRT 模型](https://github.com/emgucv/models/blob/master/mobilenet_v1_1.0_224_float_2017_11_08/mobilenet_v1_1.0_224.tflite) 2. 可在主计算机上运行以下命令： wget http://download.tensorflow.org/models/mobilenet_v1_2018_08_02/mobilenet_v1_1.0_224_quant.tgz Copy to clipboard tar -xvf mobilenet_v1_1.0_224_quant.tgz Copy to clipboard wget https://storage.googleapis.com/download.tensorflow.org/models/mobilenet_v1_1.0_224_frozen.tgz Copy to clipboard tar -xvf mobilenet_v1_1.0_224_frozen.tgz Copy to clipboard # For SCP, run the following command: ssh root@[ip-addr] mount -o remount,rw /usr cd /etc mkdir artifacts exit Copy to clipboard scp mobilenet_v1_1.0_224_quant.tflite root@[ip-addr]:/etc/artifacts scp grace_hopper.bmp root@[ip-addr]:/etc/artifacts scp mobilenet_v1_1.0_224/labels.txt root@[ip-addr]:/etc/artifacts scp mobilenet_v1_1.0_224.tflite root@[ip-addr]:/etc/artifacts Copy to clipboard 3. 要使用以下任一 delegate 运行推理，请执行以下操作： - 使用 XNNPACK delegate，在 Arm^®^ CPU 上运行模型： label_image -l /etc/artifacts/labels.txt -i /etc/artifacts/grace_hopper.bmp -m /etc/artifacts/mobilenet_v1_1.0_224_quant.tflite -c 10 -p 1 --xnnpack_delegate 1 Copy to clipboard - 使用 GPU delegate，在 Qualcomm^®^ Adreno^™^ GPU 上运行模型： label_image -l /etc/artifacts/labels.txt -i /etc/artifacts/grace_hopper.bmp -m /etc/artifacts/mobilenet_v1_1.0_224.tflite -c 10 -p 1 --gl_backend 1 Copy to clipboard > > > Note > > > GPU delegate 当前不支持原生 LiteRT 示例应用程序。 ## 后续步骤 - [部署 LiteRT 模型](https://docs.qualcomm.com/doc/80-70020-54SC/topic/tensorflow-lite-developer-workflow.html#tensorflow-lite-developer-workflow) - [运行 LiteRT 示例应用](https://docs.qualcomm.com/doc/80-70020-54SC/topic/sample-applications.html#run-litert-sample-apps) Last Published: Oct 12, 2025 [Previous Topic LiteRT 概述](https://docs.qualcomm.com/bundle/publicresource/80-70020-54SC/topics/tflite-landing-page.md) [Next Topic LiteRT 架构](https://docs.qualcomm.com/bundle/publicresource/80-70020-54SC/topics/arch.md)